Clean-up / consolidation of m68k bus error / address error handling
code. This is not a very radical change for most m68k platforms,
since they were using a version of this re-factored code already via
an #include directive in locore.s. Now, however, the bus error handlers
are split into CPU "generation"-specific files, and brought into the
kernel via the CPU options in the kernel config file (the goal being
to reduce boilerplate in each locore.s).
The more notable changes:
- amiga and atari no longer have their own copy of each handler, and
some of their local tweaks have been folded into the shared code.
- sun2's bus error handling code has been re-factored into a shared
buserr_10.s to facilitate support for other 68010-based systems (all
of which have their own custom MMU).
- sun3's bus error handling code has been re-refactored into the
shared buserr_2030.s. The Sun3-specific MMU handling is treated
just like the HP-specific MMU handling that was already there.
This additional #ifdef is a small price to pay for more code sharing.
- sun3x now also uses the shared buserr_2030.s rather than an identical
[12 lines not shown]
Import gdb-17.1, previous was 16.3
This version of GDB includes the following changes and enhancements:
* x86-64 CET shadow stack support
* Debugging Linux programs that use AArch64 Guarded Control Stacks
* GDB record feature now supports rv64gc architectures
* 'info threads' command support for two new options '-stopped' and
'-running' to limit the list of displayed threads.
* On Linux and FreeBSD, the addresses shown by the 'info
sharedlibrary' command are now for the full memory range
allocated to the shared library.
* Linux checkpoints now work with multiple inferiors
* Improved linker namespace support
* Warnings and error messages now start with an emoji (warning
sign, or cross mark) if supported by the host charset.
Configurable.
* Built-in support for TLS on Linux as backup when libthread_db is
not available. Supported on the x86_64, aarch64, ppc64, s390x,
[72 lines not shown]
libpciaccess: update to 0.19.
This release adds a new pci_device_is_boot_display() API, with support
currently implemented only for Linux systems using sysfs.
It also adds a -Dinstall-scanpci option to the meson configuration for
those who want to install this version of scanpci.
In addition, it corrects build issues on FreeBSD 15 and DragonFly, and
fixes a long-standing bug if there were multiple users of libpciaccess
in the same process and one called pci_system_cleanup before the others
were done using pciaccess.
www/chromium: update to 146.0.7680.80
* 146.0.7680.80
This update includes 1 security fixes. Please see the
Chrome Security Page for more information.
[N/A][491421267] High CVE-2026-3909: Out of bounds write in Skia.
Reported by Google Threat Analysis Group on 2026-03-10
Google is aware that an exploit for CVE-2026-3909 exists in the wild.
gpt: fix compile-time integer overflow for terabytes and petabytes
Previously, overly large values for mediasize in TB and PB were accepted
and processed modulo 1<<64.
lint complained: main.c(223): warning:
'18446744073708503040 * 1024' overflows 'unsigned long'
Improve the type builtin command slightly.
Get rid of the "is a tracked alias for" nonsense response, which
adds no useful information (just confusion) - if "type foo" said
foo is /usr/bin/foo, the next time it would say foo is a tracked
alias for /usr/bin/foo (with nothing else changing or happening
at all). I had intended to remove that attribution years ago,
and I was actually sure that I had done so, but it turns out, that
somehow it got forgotten. Now the result for "type foo" (assuming
that foo isn't changing externally, by being renamed, or created,
or something similar, or PATH changing) will be the same each time
the command is run.
While doing that, also improve the output should someone request
the type of '' (probably as a result of type "${foo}" where foo
is empty, or a name typo, or similar).
lint: do not warn about snprintb mixing 'f' with ':'
Most often, 'f' is used with '=', but in cases where the description
does not start with an identifier, there is no risk of pasting the
number and the value description together without a separator.
snprintb: allow 'F' and 'f' to be freely mixed with '=' and ':'
This restores the behavior from NetBSD 10.
While in most cases, 'f' is used with '=' and 'F' is used with ':',
there are useful cases for the other combinations as well, so allow
them.
Planned for pull-up to NetBSD 11.
Bring the mbedtls replacement mechanism up to date, just in case it
is needed.
Upstream has fiddled with the git release tags, update the
distribution tarball hashes.
gtk-gnutella: updated to 1.3.1
1.3.1
Bug Fixes
[GTK2] Removed "clock" icon in status bar, a left-over of the 1.3.0 edits.
Was sometimes crashing at startup due to a bug in logfilter.
Forgot an include of "mempcpy.h", causing compilation error on MacOs.
Fixed compilation error on arm64.
1.3.0
Improvements
Removed expiration date: this version will now run forever.
Under the Hood
[2 lines not shown]
www/chromium: update to 146.0.7680.75
* 146.0.7680.75
This update includes 2 security fixes. Please see the
Chrome Security Page for more information.
[N/A][491421267] High CVE-2026-3909: Out of bounds write in Skia.
Reported by Google on 2026-03-10
[N/A][491410818] High CVE-2026-3910: Inappropriate implementation
in V8. Reported by Google on 2026-03-10
Google is aware that exploits for both CVE-2026-3909 & CVE-2026-3910
exist in the wild.
* 146.0.7680.71
This update includes 29 security fixes. Please see the
Chrome Security Page for more information.
[$33000][483445078] Critical CVE-2026-3913: Heap buffer overflow
[62 lines not shown]
tests/snprintb: ensure that the value has power-of-two digits
In direct calls to h_snprintb_m_len, several arguments are integers, so
provide a way to distinguish them visually by forcing the value to be
given in hexadecimal. The buffer sizes are typically given in decimal.
Prevent excess or forgotten digits by requiring the number of digits in
the value to be a power of two.
audio/fasttracker2: Update to 2.11
Changes since 2.10
v2.11 - 09.03.2026
- Fixed: When loading stereo AIFF/WAV samples, the last sample point would not
be read correctly when either reading the right channel or downmixing to mono.
